Cooling Drag (technical)

[Nuisance]

To my fellow fanatics

Recently, there has been a thread about cooling drag on the Lancair email list. I have obtained the permission of the author, Fred Moreno, to post it on my website.

I think that once you 'freeze' the airframe (by that I mean use an existing design, like a Van's kitplane), the things you can do to go faster (be more efficient) are few. A straight airplane, a general cleanup, and then the main area you can work on is engine cooling drag.

It is lots of work though!

Anyway, go here

http://www.lazy8.net/speedmods.htm

and scroll down to "An Engineer's Perspective".

Have fun!

[Bob Axsom]

Well actually I am printing it as I write this message. As always - good stuff! Will study more completely. I have pulled the lower forward baffle and extensions out of my plane and will test again when I get the magneto reinstalled. My records during the evolution of my lower cowl mods show that that is the fastest configuration I have come up with so far. I'm hoping to see that 173 kt range come back. I am keeping all of the rear baffling and the horizontal baffles along the cylinders in the lower cowl and I will try the cooling air inlet plugs with that configuration.

Bob Axsom

[Kevin Horton]

Originally Posted by Bob Axsom

I have pulled the lower forward baffle and extensions out of my plane and will test again when I get the magneto reinstalled. My records during the evolution of my lower cowl mods show that that is the fastest configuration I have come up with so far. I'm hoping to see that 173 kt range come back.

Bob - Do you do multiple flights to test speed with each configuration, or just one? There are a lot of reasons why results from any one flight might be off, so if you are trying to measure small changes you really need to get results from multiple flights.

If you hope to see the effect of small mods, you also need a technique that gives results that are independent of the wind speed. Simply averaging the GPS ground speeds on different headings will give a result that depends on the wind. The wind induced error can easily be greater than the speed difference from any small mod.

[Bob Axsom]

I have finally come to recognize what you and John Huft have been saying all along. I used the US Air Race Inc. Handicap procedure for over 20 flights (I'm very stubborn) before I tried the obvious two-way straight into the wind and reverse and I also tried random two way and 4-way all using the USAR procedure for establishing the flights at 6,000 ft density altitude. My observation based on this limited sample way the multi-directional methods that were not 2-way directly into the wind and reverse yeilded a compromise speed. I find myself sneaking a peek at the TAS window of the airspeed indicator to get a clue of how I'm doing independent of the GPS techniques. The problem with the 2-way method is it is difficult to get a wind of constant direction and velocity over the Ozarks where I do my testing (or anywhere else I suspect). After trying this method a couple of times I gave up and went back to the USAR method. In the one flight where I made two random 2-way (two opposite directions), a random 4-way (two 2-ways of opposite directions perpendicular to each other), and the USAR method, the USAR yeilded the lowest number and one of the 2-ways provided the highest number.

Another error I think I have introduced into my test method is using the TruTrak Pictorial Pilot and Altrak for constant direction and altitude. My thought is I am burning true airspeed in fighting cross wind and vertical air movements and what I am recording is a resulting track speed.

In the case of multiple flights on different days with the same configuration using the USAR method I found my speeds for the tests fell across a 4 kt range.

Still I get a sense of what is working and what is not to the extent that something that I have spent many months developing and looks really good can be scrapped with a clear conscience. I wish there was a wind tunnel that I could work with to "tune" these changes under strictly controlled conditions but it continues to be a crude cut and try relatively gross change approach that is very lucky to hit the optimum.

I'm sorry for the long winded response the short answer is I do not make multiple flights to test new mods. You are right of course if I want an accurate evaluation multiple flights are required. I tend to say, well that worked or it didn't and usually it didn't.

Bob Axsom

[bumblebee]

Originally Posted by Bob Axsom

Another error I think I have introduced into my test method is using the TruTrak Pictorial Pilot and Altrak for constant direction and altitude. My thought is I am burning true airspeed in fighting cross wind and vertical air movements and what I am recording is a resulting track speed.

Bob...with all due respect, you are seriously overthinking this whole process and making it needlessly difficult on yourself.

The most accurate method for testing drag mods is to "Fly the square".

Fly a timed square (30 sec legs + turns) on the cardinals (0, 90, 180, 270) at defined power (2500/WOT) and defined DA (e.g. 8000') using shallow banks in the turns (15* or standard rate turns).

use a recording GPS and/or EFIS if available
fly 3 laps per test (minimum)
fly 3 tests per mod (ideal)
start with same conditions (full tanks, engine temps stabilized, etc.)
average all legs for GPS and TAS
calculate standard deviation (R-squared)

Using A/P and Altrak is HIGHLY recommended to ensure precision and consistency....you simply cannot hand-fly as accurately as a/p+altrak

Your results using this method should be within +/- 1% .... FAR more than enough for anything except scientific (aka divorce-worthy) tinkering.

At RV speeds, a 1% error is effectively impossible to measure or notice.

Originally Posted by Bob Axsom

In the case of multiple flights on different days with the same configuration using the USAR method I found my speeds for the tests fell across a 4 kt range.

You are doing well if you consistently achieve a 4kt range. Remember, a 4 kt RANGE = a 2 kt MARGIN (up or down)...this is very close to the minimum measurable variation.....don't get crazy trying to measure beyond this....it ain't worth it and you'll never know for sure if you measured an "improvement" or just a change in the wind. The "last knot" will be the most expensive and the least accurate.

After this, the only thing left is to bribe Paul Dye to get you into the NASA wind tunnel....and that's big bucks on both counts

[Bob Axsom]

I can't defend the approach I took as the "right" approach but I do have to maintain a healthy measure of independence and skepticism. I believe that the "square" method is a reasonable compromise method that reveals maximum speed only on dead calm days and it matters whether you fly constant heading or constant track to obtain your data. I have flown repeated cycles in the same flight and the speeds derived from individual cycles have been within one knot of each other. The four knot variations occurred when the tests were conducted on different days with the same configuration. My airplane is fairly clean and I have found it extremely difficult to gain any speed regardless of the mental, physical ot spiritual effort and knowing from personal experience that the speed I record may be off the actual maximum by many times the amount gained or lost by the modification causes the evaluation process to drift in a subjective direction. Having a disciplined approach like the square method or the triangular method or the 2-way method makes the testing more objective and I have found for me personnally that is VERY important. My autopilot will not fly a constant heading and its inflight corrections for real time variations in cross wind will always cause speed measurements to be less than the actual true air speed. The same is true of or the altitude hold function over hilly country with a rising and falling air mass. Test flights early in the mornings of calm days provide the most reliable results for me but when I'm ready and able to test, I test, grudgingly accept the results, try to rationalize what they mean and decide what to do in the future. For the professional pilot or the everyday pleasure flier you take what you get and that is that but if you are modifying your own plane with a desire to find the best configuration for racing speed and you have a demonstrated four knot error range you can't just throw the data on the table and walk away. You have to eventually decide whether the mod stays as is, gets modified or removed. It is not easy! This is the kind of problem that leads the test organizations to develop the "fly the squares and average the results" methods I'm sure. Testing costs money and schedule so procedures are essential to control this imperfect activity. ARGH! my head hurts I'm going back to bed.

Bob Axsom

[Kevin Horton]

Originally Posted by bumblebee

Bob...with all due respect, you are seriously overthinking this whole process and making it needlessly difficult on yourself.

The most accurate method for testing drag mods is to "Fly the square".

Fly a timed square (30 sec legs + turns) on the cardinals (0, 90, 180, 270) at defined power (2500/WOT) and defined DA (e.g. 8000') using shallow banks in the turns (15* or standard rate turns).

use a recording GPS and/or EFIS if available
fly 3 laps per test (minimum)
fly 3 tests per mod (ideal)
start with same conditions (full tanks, engine temps stabilized, etc.)
average all legs for GPS and TAS
calculate standard deviation (R-squared)

Simply averaging the GPS ground speeds does not fully account for wind. For example, if our TAS is 160 kt, and the wind is 045 degrees at 25 kt, our ground speeds will be:

hdg GS
000 143
090 143
180 179
360 179

average = 161

if the wind is 50 kt (not at all uncommon at altitude), the average ground is up to 164.5, with a TAS of 160 kt. This isn't nearly accurate enough to see the effect of a small mod.

If you are going to go to the trouble to fly the tests, you might as well use a calculation technique that gives an answer that is independent of wind. I highly recommend the NTPS Spreadsheet, using the four leg option.

[Tom Martin]

Over the years I have done a lot of "testing" with my rockets. Getting accurate consistant data is very difficult. I use a three leg gps system that I borrowed from Kevin Horton's web site. It is pretty good but I have found that there are huge differences in HP based on temperature so that really needs to be calculated as well. As Bob Axom has stated "it makes my head hurt"! The best most accurate testing for drag that you can do is fly beside a friend who has not made any changes. Pick power settings where you are flying exactly the same speed. Make your modification and fly the same power settings again. This takes a whole bunch of variables out of the equation and has proven to be most effective for me when I am trying to show performance changes. Last fall I flew beside my friend and got some base numbers, then changed exhaust systems, and flew again. There was no difference, nada, nothing. However a few changes to my cowling outlet area and I could show a speed increase related to the other aircraft. For me this has proven to be the simplest and most accruate way to test for speed changes.

[captainron]

I'm starting to wonder if a different thought pattern is required here. Looking at my fuselage that has yet to have the engine mount installed makes me think about the large frontal area here that needs to be covered with something. Sure, it's going to get an engine mount and then a very flat engine. And then a huge cowling just to blend in with the firewall pattern.
Thinking about all the twins I have flown and their extremely tightly cowled engines, engines which were much larger (in displacement) than what we're working with here, has me thinking that maybe a different approach is needed. Take a look at an Aerostar or a C-310 and notice that the cowling or nacelle is not much bigger than the width and height of the engine itself. Then, think about the fact that there may be TWO turbochargers, an intercooler, the rest of the exhaust system, and every other accessory found on any engine packed in there!
Our firewalls look like they have room for a radial engine to be mounted in front of them!
Don't know what the answer is except to move the engine forward enough so that the cowling can be reduced in height and volume.
How have the planes that have their engines mounted where the fuselage has to be fat anyway (E-Z's, Velocity, etc.) fared in speed compared to the Vans planes with the same HP/weight ratios?

[David-aviator]

Originally Posted by captainron

I'm starting to wonder if a different thought pattern is required here. Looking at my fuselage that has yet to have the engine mount installed makes me think about the large frontal area here that needs to be covered with something. Sure, it's going to get an engine mount and then a very flat engine. And then a huge cowling just to blend in with the firewall pattern.....
.....How have the planes that have their engines mounted where the fuselage has to be fat anyway (E-Z's, Velocity, etc.) fared in speed compared to the Vans planes with the same HP/weight ratios?

Except for one or two highly modifid airplanes, the pushers are not much better. I built a LEZ and Cozy and felt there was much turbulence across the prop that defeated some of the reduced drag of the configuration. I know there was much experimentation with positioning the prop aft of the engine to minumize this issue.

Thanks much to the guy who started this thread. I believe I have some serious drag issues to address with a modified cowling to accomodate the Subaru cooling needs and this is will help sorting it all out.